The present invention relates to an image processing method and apparatus for converting multivalued image data to data corresponding to each recording element, and an image forming apparatus for inputting multivalued image data and forming an image.
Recently, as personal computers are widely used, various types of printers are used by general users. In these printers, there are increasing demands for high-speed printing, high resolution, high image quality, low noise and so forth. To satisfy such needs, there are printers employing the ink-jet printing method. Because of the relatively small size and low noise, printers employing the ink-jet printing method have rapidly become popular.
Many of the ink-jet printers utilize a printhead where a plurality of ink discharge orifices (printing elements) are integrally arrayed to improve printing speed, or utilize a plurality of printheads for color printing. Furthermore, to satisfy the needs for high resolution and high image quality, these ink-jet printers perform half-tone processing such as a dither method or an error diffusion method, as a method to faithfully reproduce tones of an original image. Based on image data on which such half-tone processing has been performed, printing is performed.
When printing is performed by print data processed by the above-described image processing method, if the resolution of a printing apparatus is high (1000 dot per inch (dpi) or more), supreme tone reproduction is possible. However, if the resolution of a printing apparatus is low (360 dpi to 720 dpi), dots in the highlighted portions become conspicuous, resulting in image coarseness due to non-uniformity of pixels. In order to increase the number of tones, the print dot itself is expressed in multivalues.
As a method of improving tone representation and obtaining an image having high density and wide range of tones by an ink-jet printer, a number of printing methods are proposed and put into practice. One is called a multi-droplet method where one dot is formed by discharging a plurality of ink droplets on substantially the same position of a print medium and tones of the dot are expressed by changing the number of ink droplets discharged on the dot. Another method utilizes plural types of ink having different densities. Tones of a dot are reproduced by combining at least two types of ink droplets having different densities of the same color group. Alternatively, a printing method combining the above-described two methods is also proposed and put into practice.
As a method of pseudo-tone reproduction, there is an error diffusion method (reference: R. Floyd & L. Steinberg, "An Adaptive Algorithm for Spatial Gray Scale" S1 D75 Digest, pp. 36-37). As an example of an image printing method utilizing the error diffusion method, "Multivalued Error Diffusion Method" by Katoh, Y. Arai, and Y. Yasuda (National Conference of Communication, Department in Showa 53 year, Society of Electronic Communication in Japan (1973), pp. 504 (Japanese)) is known. This method utilizes an error diffusion method utilizing a plurality of threshold values, whereas the conventional error diffusion method utilizes one fixed threshold value. For instance, assume that the range of image data is 0 to 255. Conventionally, "128" is set as a threshold value and the conventional error diffusion is performed to obtain binary data. However, according to the "Multivalued Error Diffusion Method" by Katoh, Y. Arai, and Y. Yasuda, in a case of using two types of ink having different densities, "85" and "175" are set as threshold values and ternary data is obtained for using the two types of ink in accordance with the print density.
Furthermore, an attempt has also been made recently to reproduce a high-quality image by obtaining multivalued data with three or more types of printing densities. For example, there is a method of using a combination of ink and film having an additive property to print a transmission image with multivalued dots. Herein, when the combination of ink and film achieves an additive property in printing a transmission image on a film by an ink-jet printing method, density is added up as ink droplets are superimposed plural times on the same pixel position (this density will be referred to as transmission density). An example of achieving an additive property is described hereinafter.
BJ transparent film CF-301 (manufactured by CANON) is used as a print sheet. On this film, dots are uniformly printed using 2% solution of a dye-type ink, C.I. Direct Black 19. As a result, an image having a transmission density of "0.8D" is obtained. Similarly, dots are uniformly printed using 1% solution of the C.I. Direct Black 19. As a result, an image having a transmission density of "0.4D" is obtained. If printing is performed by superimposing dots of these two types of ink, an image having a transmission density of "1.2D" can be obtained. It is confirmed from this experiment that the combination of the ink and film substantially obtains the transmission density in the range of "0" to "2.5D".
In a case of using such combination of the ink and film which achieves the additive property, the number of tones reproducible can be significantly increased by superimposing plural types of ink having different densities.
However, when printing is performed with the aforementioned image processing, inputted multivalued image data needs to be converted to binary data in correspondence with each ink-jet head (type of ink) The inputted multivalued image data is stored as multivalued data (e.g., 8 bits) in each address of an image memory. After the multivalued image data is developed into binary data by the error diffusion method or the like, the binary data is stored in a memory (bit plane memory) different from the aforementioned image memory, which corresponds to the type of ink or the number of ink-jet heads.
As described above, conventionally an image memory and a bit plane memory are separately constructed. Thus, if the types of ink or the number of ink-jet heads are to be increased, the number or the capacity of the bit plane memory must be increased for storing binary data.